We all know that organisations are not short of technology. Their cloud and automation are well established, AI has demonstrated its potential and we can create secure environments in hours, analyse legacy code and generate tests rapidly.
Yet modernisation still takes too long.
What that tells us is that the constraint is no longer always technology, increasingly, it is the operating model around it. We have all been in meetings where everyone agrees something must change, followed by three months discussing the PowerPoint about it. I have seen organisations wait six months to approve six days of delivery. That suggests the bottleneck is not engineering.
The infrastructure is not waiting, the organisation is. And this makes this a productivity problem as much as a technology problem. When skilled people are spending more time waiting for decisions and navigating approvals over delivering change, modernisation becomes more expensive without being more effective.
Stop proving what we know
I’m increasingly sceptical of pilots designed to prove mature technology – another controlled demonstration is unlikely to reveal anything fundamentally new.
Let’s say that a pilot succeeds through careful selection, exceptional attention and protection from organisational difficulty. Integrations are limited, security is accommodating and specialists surround the work. The pilot succeeds – but what has really been proved?
Can it operate across the estate, meet assurance and outlast the original team? Can it scale from one application to 50 and sustain value?
A proof of concept proves a concept; it doesn’t prove that an organisation can transform. Pilots still have a role where the technology or hypothesis is uncertain, but we should be clear about the uncertainty being removed. A better alternative is a value-bearing first move: bounded enough to manage safely, but real enough to improve an important service or outcome. This move should encounter genuine constraints, improve the estate and create a reusable pattern and evidence for the next decision.
The opposite of a big-bang programme is not a pilot. It is controlled, incremental execution.
An inventory is not an understanding of the estate
Most organisations can produce an inventory. They can list applications, servers, cloud resources, databases, suppliers and contracts. But an inventory shows what exists, not necessarily what matters. The estate is defined by the relationships between them.
Which application feeds critical services? Which unsupported component underpins a modern front end? Which supplier holds undocumented knowledge?
Legacy rarely announces itself with a large red warning light. It usually hides in a dependency nobody recorded. And the most dangerous system may not be the oldest, it may be the one nobody realises everything else depends on.
Age is a poor basis for investment. Some older systems remain stable and valuable; newer ones may create drag because they are expensive, poorly integrated or difficult to change. A better approach is to ask whether the technology creates a constraint. Does it prevent service improvement or create unacceptable security, resilience or supportability risk? Is every change disproportionately slow and expensive? Does it consume scarce capacity through incidents and workarounds?
Modernisation should target the constraint, not the birthday.
Progressive confidence, not perfect knowledge
The answer is not another endless discovery. Technology is dynamic: applications, risks and priorities change. By the time a perfect model is complete, part may already be outdated. The objective should be progressive confidence where you understand enough to make the next responsible decision. You make a bounded intervention under real conditions, then use delivery evidence to improve the next decision.
That does not mean ignoring uncertainty of course. Assumptions and gaps should be visible, but incomplete information should not excuse inaction. You do not need to understand everything before doing anything. You need to understand enough to do the next thing responsibly.
Insight also has a shelf life. If it does not change a decision, it is already depreciating.
Move governance from documents to decisions
Good governance is not the enemy of speed like many believe. Without good governance comes duplicated investment, conflicting decisions, unmanaged risk and solutions that cannot scale. Governance should accelerate routine decisions and focus leaders on exceptions.
A six-month approval process does not become modern because the infrastructure at the end of it is cloud. The fastest organisations move governance closer to decisions: clarifying rights, letting teams work within guardrails, reusing proven patterns and escalating exceptions.
The question should not be: “Which document is missing?” It should be: “What decision has not yet been made?”
Modernisation accelerates when organisations stop managing document queues and start managing decisions, constraints and evidence.
Cloud changes the economics of change
The value of cloud is not simply a different place to run technology. Its greater value is the potential to change the economics of making the next change. Reusable foundations, standard patterns, automated controls, deployment pipelines and retained assurance evidence make change faster, safer and repeatable. The test is whether the second change becomes easier than the first, and the tenth easier still.
If every deployment remains bespoke, manual and dependent on a few specialists, an organisation may have adopted cloud technology without cloud economics.
AI will expose what modernisation has left unresolved
AI ambition eventually meets the reality of the estate. Fragmented applications, inaccessible data, hidden business rules, undocumented interfaces and unclear ownership do not disappear when AI is introduced. Putting AI on top of fragmentation does not remove complexity. It gives complexity a conversational interface. AI readiness is not simply model selection; it is also a modernisation, data and operating-model challenge.
Transformation is a capability
Public-sector organisations face rising expectations alongside constrained funding, skills and capacity. Improving productivity therefore can’t simply mean asking people to do more. It has to mean removing the constraints that prevent people and technology from delivering value. That makes prioritisation unavoidable because almost everything matters to someone. The challenge however, is deciding what matters most now.
Modernisation should not merely produce a newer estate, it should leave the organisation better able to make the next change. And that requires connecting evidence, decisions, execution, assurance and outcomes while learning and acting without losing accountability.
You must stop proving that cloud works. You should start proving that transformation works.
Make the first move count. Modernise genuine constraints, then build the capability to understand, decide, execute and learn continuously. Because the strongest transformation does not simply produce better technology, it produces an organisation that is better at changing it.